About Me

My Photo

I previously worked on Virtual Reality and other hardware at Valve.  I currently work at Google[x].

Prior to starting at Valve, I built computer peripherals such as keyboards, mice, and joysticks that were designed to be used inside MRI machines.  My company, Mag Design and Engineering, sold these devices directly to researchers at academic institutions who used them to publish scientific papers in peer-reviewed journals.

After work, I spend time on many different types of projects that usually involve circuit design, machining, material selection, and general fabrication/hacking.  My favorite place to be is my home workshop.

ben dot krasnow at gmail

http://www.youtube.com/user/bkraz333

http://www.linkedin.com/pub/ben-krasnow/4/6a9/679

http://www.twitter.com/BenKrasnow

Sunday, April 19, 2009

DIY aquarium chiller is a success

Today was the first warm day of the season here in Sunnyvale, CA. In my living room, the air temperature was close to 90*F


The aquarium temperature was a steady 80*F throughout the day.
The temperature controller claimed to be using only 10-15% of the chiller's capacity, but I think this might be misleading because of the way the system is setup. The chiller (water cooler) has its own thermostat, and attempts to keep its water temperature around 45*F. The fishtank's temperature controller turns a pump on and off that pushes the chilled water through a heat exchanger with the aquarium water. The amount of time that the pump is running is the "percentage of capacity" that I have been listing here and in other posts. I am sure the system has a non-linear response such that the amount of cooling delivered at %100 would not be ten times the cooling delivered at %10. This is because the chiller's water temperature would be rising (making it less effective at cooling the tank) as the pump runs more often. Nonetheless, I think the chiller system has plenty of headroom, and it only rarely gets hotter than 90*F in my house, so I am feeling pretty good about the project.

Wednesday, April 8, 2009

Monster TIG nozzle = monster waste of money

UPDATE 6/15/2013: Arc-zone has redesigned their Monster TIG nozzle. It now has a chunk of rock wool or similar material to help diffuse the gas flow. I have never used this version, and so the comments in this blog post do not apply to it. Check the comments section for more details.


Today's lesson involves my quest to weld stainless steel sheet metal, a Monster TIG nozzle, and a copper chill bar. In previous posts, I've described having trouble maintaining weld bead quality on thin stainless sheets. The problems are a combination of putting too much heat into the metal, and having too little argon gas coverage. I am not sure if addressing one problem can help solve the other. Today, I did some testing to find out if adding a lot of gas coverage can help. I also tested out a copper chill bar.

Common weld parameters for the whole test:
1/16 ceriated tungsten ground to a sharp point
55 amps (pedal floored for the entire test)
very slight %90 pulse at 200Hz just to get my auto-darkening helmet to work
20 CFH pure argon
10 sec post-flow

I purposefully used a small piece of 304 1/16" sheet to show the heat buildup problems. I also welded close to the edge to test the worst-case heat buildup.

First up: normal gas lens with #8 cup.

Wow, I never knew the copper could help that much!

Next, a large diameter gas lens with #12 cup
Same story here. It looks like there was even less heat in the metal. This might be because there was better contact between the sheet and copper, or because the gas nozzle has a wider opening. I'll bet the Monster nozzle will be even better...

Finally, the "Monster TIG nozzle", which is 1" in diameter and uses a stubby gas lens collet body.
Wha?! There must be something wrong -- what's going on here?! I tried all gas flow settings from 5 CFH up to 30 CFH and concluded this nozzle is completely useless. It's possible that I am misunderstanding something since I am a new welder, but I am pretty sure this thing just plain doesn't work. I noticed that the tungsten had turned black after a few welds, indicating the gas coverage isn't even enough to keep the tunsten from oxidizing. At 30 CFH, the gas flow was so turbulent, I could see the arc getting blown around, and pops of smoke coming out of the weld. At lower flow settings, I could see the stainless oxidizing even before I lifted my hood. I tried different stickout from 1/8" up to 3/4" with no change. I am sure the cup made good contact with the torch body, and there were no air leaks. I even tried extending the nozzle away from the gas lens with a spacer to make sure there was adquate space for the gas to disperse with only a tiny improvement.

On the right: large gas lens with #12 cup. On the left, you guessed it, Monster suck.



This screen arrangement doesn't look so great.

It's made with just two screens without any spacers between them, and two very coarse screens on the outsides. The screen diameter is a few mm less than the interior diameter of the ceramic cup, so I'm guessing a lot of gas slips around the edges of the screens.
So, I'll be continuing my stainless welding quest without the Monster nozzle and with copper chill blocks. I'll also be testing Solar Flux B. So far, I think it works well but poses a huge cleanup mess after the welding is complete.

Saturday, April 4, 2009

Even more detail for aquarium DIY top-off system

In this post, I'll describe the machines and methods that I used to make the optical sensor head for the aquarium top-off system.

Step 1: Cut off 2" of 1/2" diameter black Delrin rod. Delrin is a brand name for a specific type of acetal copolymer plastic. You can get it at www.mcmaster.com


Step 2: Square-off the end of the rod on the lathe.


Step 3: Drill two holes at 30* to the rod's major axis so that they intersect a few mm in front of the rod's face. You may need to use a punch or dremel to prevent the drill bit from wandering on the sloped face of rod. I like to use the vise-within-a-vise method for drilling angled holes on the drill press.

Step 4: Taper the end of the rod with a coarse file.

Step 5: Use a smooth file to tidy everything up.

I found that the optic fibers should be cut square, and then recessed into the sensor head so that nothing sticks out. This allows air bubbles to form harmlessly in each of the spaces created by recessed fiber ends. If the fibers stick out too much, an air bubble might become trapped between them, and the system will have a "false positive".

Thursday, April 2, 2009

More detail for the DIY auto top-off system for aquariums


I've had a few requests for plans of the auto top-off system for aquariums. This project is fairly straight-forward, and I would imagine the most difficult part is machining the sensor head that holds the two optical fibers. This will require a drill bit that is sized just right, and the angles need to be pretty accurate. The wiring and electrical side of this project is very easy, and only requires minimal soldering. Here's some rough instructions and part list that I emailed to an interested DIYer:

This project involves wiring a household electrical outlet. As you know, saltwater is a great conductor and the wiring must be protected from drips, etc. Always disconnect the power when working on the circuit, and use good wiring practices.

I'd start by getting a Keyence fiberoptic sensor FS-V11 (or similar) ebay like this:

http://cgi.ebay.com/Keyence-FS-V11-FSV11-Fiber-Optic-Sensor_W0QQitemZ390020341450QQcmdZViewItemQQptZLH_DefaultDomain_0?hash=item390020341450&_trksid=p3286.m20.l1116

Next, you need a wall-wart power supply (AC adapter) to provide low voltage to the Keyence. You'll need 12V DC at almost any current rating (mA rating). Be sure the AC adapter is a linear voltage supply, not a "switching" supply -- it should be relatively heavy.

You'll need some 1mm jacketed plastic (PMMA) fibertoptics. I didn't find any at a great price on eBay. You'll have to hunt around a little. Edmund Optics sells it per foot, so you might be able to just order a small amount.

Get the Aqualifter AW20 pump ($10 or $15)

Get a solid state relay eg Kyotto KB20C02A (Jameco #175214) $6.55

You can wire it all together in a standard electrical box with an electrical outlet from a home improvement store.

The basic idea is that the Keyence device gets power from the 12V DC supply, then controls the solid state relay with its output. The solid state relay controls power to the electrical outlet, which is where the AW20 pump is connected.

I don't know if I would recommend this project to someone who has never done any electrical/electronics projects before. It involves wiring a household electrical outlet, and this really has the potential to cause a lethal shock -- especially around saltwater aquariums.

Another problem is that a malfunction might flood your house! I noted problems about air bubbles. If the fiberoptic sensor head traps a large air bubble, the top-off pump will continue running even after the aquarium overflows. For the first week or two, I used the top-off system as just an indicator -- not risking any floods. You should do the same if you attempt this project. I modified the sensor head, so that I doubt air bubbles will cause any more problems, but you never know.

Feel free to give me more feedback in the comments...

Wednesday, April 1, 2009

Lathe-turned wood vase

Here is another project that I completed a couple years ago. It was a wedding present for a good friend.

I started by laminating pieces of rosewood, padauk, maple and walnut in a symmetric pattern. After the glue dried, I sawed off the corners, and mounted it in my wood lathe.
This was the largest chunk of wood that I had ever turned. The lathe took almost 5 seconds to get up to speed, and I was more than a little concerned about the wood coming loose.


I only have one proper wood lathe tool. It's a rough gouge, and I use only that tool to get very close to the final shape on most of my projects.

I use chisels, files, rasps, and finally sandpaper to finish the piece. I usually sand to 220 grit while the piece is still on the lathe. For the final pass, I stop the lathe and sand by hand in the direction of the grain.

I built a crazy-long extension for a Forstner bit.


I made an acrylic tube that sits down into the vase, allowing it to hold water without damage to the wood. The top section of the vase is painted with black semi-gloss. The wood finish is either wipe-on polyurethane or a Tung oil finish (I can't remember).